无刷直流电机输出转矩的非线性偏最小二乘回归估计

发布时间：2018-01-15 02:25

本文关键词：无刷直流电机输出转矩的非线性偏最小二乘回归估计　出处：《电工技术学报》2017年05期 　论文类型：期刊论文

【摘要】：针对无刷直流电机直接转矩控制系统中,电机输出转矩估计或观测存在需检测电机状态信号难以精确地检测、观测或计算等不可避免的难题,提出一种利用电机线电流和转速两个易检测信号进行电机输出转矩估计的方法。该方法首先基于拉丁超立方抽样方法确定建模数据,确保离散数据的可靠性;然后通过核函数将建模数据中的自变量变换到特征空间,将自变量与因变量之间的非线性关系在特征空间中线性地展开;最后在特征空间中利用偏最小二乘回归方法建立电机输出转矩的估计模型。通过对转矩估计精度和控制系统性能进行分析及实验验证,证明了所提转矩估计方法能准确地估计电机输出转矩,确保电机直接转矩控制系统的性能,该方法具有较强的实用性。
[Abstract]:In the direct torque control system of brushless DC motor, there are some unavoidable problems such as estimating or observing the output torque of the motor, which is difficult to detect, observe or calculate the motor state signal accurately. This paper presents a method to estimate the output torque of the motor by using two easily detectable signals: linear current and rotational speed. Firstly, the modeling data are determined based on Latin hypercube sampling to ensure the reliability of discrete data. Then the independent variables in the modeling data are transformed into the feature space by the kernel function, and the nonlinear relations between the independent variables and the dependent variables are expanded linearly in the feature space. Finally, the estimation model of motor output torque is established by using partial least square regression method in the feature space. The precision of torque estimation and the performance of control system are analyzed and verified by experiments. It is proved that the proposed torque estimation method can accurately estimate the output torque of the motor and ensure the performance of the motor direct torque control system.
【作者单位】：道路灾变防治及交通安全教育部工程研究中心;工程车辆安全性设计与可靠性技术湖南省重点实验室;长沙理工大学汽车与机械工程学院;
【基金】：国家自然科学基金项目(51675057) 长沙理工大学道路灾变防治及交通安全教育部工程研究中心开放基金项目(kfj110304) 交通运输部应用基础研究项目(2013319825090)资助
【分类号】：TM33
【正文快照】： 0引言无刷直流电机(Brushless DC Motor,BLDCM)具有能量密度大、控制简单、运行稳定、调速性能好等优点,已被广泛应用于各领域[1]。为了提高控制系统的动态特性,降低转矩波动,近年来无刷直流电机的直接转矩控制(Directed Torque Control,DTC)已成为电机控制研究的重点内容。注

【相似文献】